Follow-up Comment #1, bug #68897 (group groff):

Dickson's analysis:

> When a .T& continued format fails the "last row of format is all lines"
> check, process_format deletes the format at main.cpp:1317. That object is
> still owned by process_table, which deletes it again at line 1685. Four lines
> of tbl input are enough:
> 

> .TS
> l.
> .T&
> =.


> 
> I reproduced it on the 1.25.0 tarball (same main.cpp as master cc834ee9)
> under ASan, where it's a heap-use-after-free in format::~format(). Stock
> Ubuntu 24.04 groff 1.23.0 segfaults on it too, and man -l on a page with this
> table ends in "man: tbl: Segmentation fault". I only saw a crash, no
> controlled write, so I'm not claiming more than that.
> 
> A two-line fix (only delete when current_format is null) makes the ASan run
> clean, and a normal .T& table renders the same. Everything is in this gist-
> input, man page, patch and my output:
> https://www.google.com/url?q=https://gist.github.com/DicksonWu654/7bdd1fc80e137d6a96efe85d41d30203&source=gmail&ust=1794166745311000&sa=E
> 
> Happy to test a different fix if you'd rather do it another way


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?68897>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/

Attachment: signature.asc
Description: PGP signature

Reply via email to